The Six-Month Autonomy Razor
If you're still telling a hire what to do six months in, you hired the wrong person.
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 3
- Confidence
- 95%
Deng's core hiring and management test: he optimizes for people who, within six months, are telling him what needs to be done rather than the reverse. The razor simultaneously keeps his hiring bar high, communicates the standard to new hires, and reframes the manager relationship around building autonomy instead of hitting individual OKRs.
Origin
Peter Deng's personal operating principle, documented in his 'PXD API' doc ('how to work with me') carried across companies.
Core principles
- 01A manager's job is to scale; without autonomous reports you cannot build the best product
- 02Every mistake becomes a learning opportunity when the shared goal is calibration, not OKR compliance
- 03The razor pressures both sides — the manager to set the person up, the report to grow into ownership
- 04Applies to every manager, not just senior leaders — it is how you pass on institutional wisdom and gain leverage
How to run it
- 1
Set the bar before hiring
Use the razor as a filter: ask whether, in six months, you could truly let this person run without direction. If not, don't settle.
Pro tip In interviews, look for behavioral evidence they could reach full autonomy in six months — it sharpens your picking, not just their development.
Watch out Settling on a candidate who fails this test guarantees you'll still be directing them months later.
- 2
State the standard on day one
Tell the new hire explicitly: 'in six months if I'm telling you what to do I've hired the wrong person,' so they know what success looks like and have something to work toward.
- 3
Shift the meta-goal to calibration
Replace 'did you hit this OKR' with 'are we calibrating enough that you become the one telling me what needs to be done.' Judge the relationship by growth toward autonomy.
Pro tip This framing turns every misstep into a joint learning moment rather than a performance failure.
In the wild
Deng describes Joanne Jen at OpenAI as uniquely combining technical depth with product taste — 'she was telling me what to do from like day two and I loved it.' He nudged her to write up the strengths she got excited about.
→ That artifact led Deng to create and hire for a brand-new 'model designer' role at OpenAI, staffed by their first two model designers.
Common mistakes
Treating hires as warm bodies to execute your instructions
If you optimize for people who follow direction, you cap your own leverage and can never scale the team's expertise beyond yourself.
Measuring the relationship purely by OKR hits
When the goal is only whether a target was met, mistakes become failures instead of calibration data, and the report never grows into the autonomy the razor demands.
Is it for you?
Best for
Any manager or founder hiring reports they need to eventually operate independently
Not ideal for
Roles that are genuinely execution-only or highly proceduralized where autonomous redirection isn't wanted
From the transcript
“in six months if I'm telling you what to do I've hired the wrong person”
“the meta goal becomes hey are we building you know are we calibrating enough are we actually getting to a spot where in six months…”
“I think this is for everyone. I think it's for everyone who is a manager”
From the episode
From ChatGPT to Instagram to Uber: The quiet architect behind the world’s most popular products
Peter Deng